[sdiy] quadature oscillator

[harrybissell]

ASA Sadegh

The best way is to configure the OTA for a ground-referenced
resistor.  The "floating" resistor circuit is hard to make work
imho.

Bias current sould be pretty easy to do in most cases.

Conventional wisdom puts a 10K resistor in series with
the Gm bias pin... so that if you screw up the current is limited
somewhat.  If you apply a voltage source to the Gm bias, its dead.

The test is... the Gm bias pin should always sit near the negative
rail (except for the oddball configuration where they vary the negative
supply to the chip to vary the bias)

Gm should be about two diode drops above V-

I think your test method is not correct.  The OTA output is essentially
current. You rpobably want to make a voltage divider by setting your
fixed resistor to the positive or negative rails... and assume the OTA
side will go to ground... and measure that way.

If you run the OTA output to a virtual ground point (like in a state
variable filter... or dast I say it... its likely that the quadrature
oscillator has a similar archetecture :^)... then all you need to do is
use the OTA directly and forget the variable resistor circuit.

In an opamp circuit (to the inverting input) the resistors essentially
make the input voltages look like currents.  Feed an OTA to that same
virtual ground and ~voila~ .... instant variable resistor.

I'd look at some SVF circuits, then try to bend one into the quadrature
oscillator. They are near cousins. You have the quad oscillator
schematic...
so get some SVF and put them together in the blender. Shake well until
mixed :^P

Some SVF should be found if you Google for "ARPtech"  (all old ARP
schematics)
or look for "Aries" (the same designer whop did the ARP, Dennis Colin...
did
the Aries filter) or the WASP filter  (think you guys have that)

BTW... varying currents in a circuit allows a much wider range than
varying resistance in most cases.

Using the schematics from the LM13600 / 13700 data sheets for variable
resistance is something of a pipe dream.  If it worked well... we'd
have done it and long ago. The voltage controlled variable resistor is
high on the 'holy grail' list.  Often in app notes they put things to
show how
'versatile' a chip is... even if they don't really work all that well...

H^) harry


sadiq sadiq wrote:
> 
> Hello,
> 
> As part of Dr. Lanterman's ECE 4803 (Music
> synthesizers)class we are  building the quadature
> oscillator from the 074 op-amp datasheet.
> 
> We are trying to replace some resistors in an
> oscillator circuit with voltage controled resistors
> using the LM 3700 OTA.  What is the best way to do
> this?
> 
> My attempts to make a voltage controlled resistor have
> failed.  Reasons for this could be that I am
> improperly supplying the bias current( using rene
> smitzhsh circuit as the bias current) to the OTA.  Or
> the circuit for the OTA as a voltage controlled
> resistor is built wrong.  Or the
> chip is dead.  I think I have wired my circuit
> correctly, how would i be able to test or know if I
> fried the OTA.
> 
> Right now I have just been testing the OTA circuit to
> see if it functions as a variable resistor.  I have
> done this by connecting a load resistor to the output
> and common, and using it to measure the resistance
> looking into the OTA from the output.  Is this the
> wrong method for doing this?
> 
> I am doing this to determine if the circuit could be
> set equal to a resistor value, and swap it with one of
> the resistors in the oscillator circuit.
> 
> Any help is greatly appreciated
> 
> Regards,
> Sadegh
> 
> 
> 
> __________________________________________________
> Do You Yahoo!?
> Tired of spam?  Yahoo! Mail has the best spam protection around
> http://mail.yahoo.com